Viral protein hijacks a metabolic enzyme to drive Kaposi’s sarcoma spread

Kaposi’s sarcoma is a multicentric tumor of lymphatic endothelial origin caused by Kaposi’s sarcoma-associated herpesvirus, or KSHV, and its progression from indolent skin lesions to aggressive visceral disease has long resisted mechanistic explanation. A new study published in Experimental & Molecular Medicine now reveals how a single viral protein rewires the metabolism of infected lymphatic endothelial cells to fuel tumor dissemination. The research, led by Yeong Jun Kim and Hye-Ra Lee of Korea University together with colleagues at the University of Southern California, the University of Florida, Sungkyunkwan University, Wuhan University and the University of Ulsan, identifies a direct interaction between the viral interferon regulatory factor 3, known as vIRF3, and the glycolytic enzyme pyruvate kinase M2, or PKM2, as a decisive driver of Kaposi’s sarcoma malignancy.

The team began by searching for cellular partners of vIRF3 in lymphatic endothelial cells using immunoprecipitation followed by mass spectrometry. Among the proteins that specifically co-purified with the viral factor was a polypeptide of roughly sixty kilodaltons corresponding to pyruvate kinase M. Although vIRF3 could bind both the PKM1 and PKM2 isoforms in pull-down assays, only PKM2 was expressed in lymphatic endothelial cells, so the investigators focused their attention on the vIRF3–PKM2 axis. Detailed binding experiments mapped the interaction to a short region of vIRF3 spanning amino acids 250 to 280. Deleting this segment produced a mutant viral protein, vIRF3-MT, that lost the ability to bind PKM2 while retaining its interactions with other known partners, including IRF7, HIF1α and HDAC5, making it a clean genetic tool for isolating the metabolic consequences of the interaction.

The biological significance of the binding became apparent when the researchers examined the oligomeric state of PKM2. This enzyme exists in two functionally distinct forms: a tetramer that acts as an active pyruvate kinase, channeling carbon from phosphoenolpyruvate into pyruvate and onward into the tricarboxylic acid cycle, and a nuclear dimer that functions as a transcriptional co-factor supporting aerobic glycolysis. When wild-type vIRF3 was expressed in lymphatic endothelial cells, cross-linking experiments showed a striking increase in PKM2 tetramerization, accompanied by elevated pyruvate kinase activity, higher intracellular acetyl coenzyme A levels and reduced lactate concentrations. The mutant viral protein, lacking the PKM2-binding region, produced none of these effects. Seahorse metabolic flux analysis reinforced the picture: cells expressing wild-type vIRF3 showed significantly increased oxygen consumption rates, including higher basal respiration, ATP production and maximal respiratory capacity, while their extracellular acidification rate, a proxy for glycolysis, declined.

To trace where the newly abundant acetyl-CoA came from, the team performed liquid chromatography–mass spectrometry metabolomics and isotope tracing with uniformly labeled carbon-13 glucose. In cells expressing wild-type vIRF3, pyruvate accumulated predominantly in the M+3 isotopologue form, a signature of enhanced PKM2 activity, and the labeled carbon flowed into M+1 and M+2 acetyl-CoA, which then fed the biosynthesis of citrate, α-ketoglutarate, succinate, fumarate and malate. Glutamate labeling patterns indicated that the cells were not relying on glutaminolysis to replenish the cycle, and measurements of fatty acid oxidation intermediates showed that this pathway was actually suppressed: carnitine accumulated while expression of carnitine palmitoyltransferase 1A and downstream β-oxidation intermediates fell. The evidence therefore pointed to enhanced pyruvate catabolism, rather than fatty acid oxidation or glutamine metabolism, as the principal source of the expanded acetyl-CoA pool.

Acetyl-CoA is not merely a fuel; it donates acetyl groups for the lysine acetylation of thousands of proteins, including transcription factors. Consistent with this, wild-type vIRF3 expression raised the global acetylation level of cellular proteins, and this increase depended on citrate exported from mitochondria and processed by ATP citrate lyase, since the ACLY inhibitor BMS-303141 abolished the effect. A human protein acetylation microarray then identified twenty-four proteins whose acetylation rose specifically in the presence of the vIRF3–PKM2 interaction. Among them, the investigators prioritized SMAD5, a transcriptional effector of TGFβ–BMP signaling with established roles in endothelial-to-mesenchymal transition, the process by which endothelial cells lose their identity and acquire invasive, mesenchymal characteristics.

Immunoprecipitation confirmed that SMAD5 acetylation was markedly increased in cells expressing wild-type vIRF3 but not the binding-defective mutant, and co-immunoprecipitation identified the acetyltransferase p300 as the enzyme responsible, since the p300 inhibitor A-485 reduced SMAD5 acetylation. The functional consequences were dramatic. Cells expressing wild-type vIRF3 adopted a spindle-shaped, mesenchymal morphology, downregulated the endothelial markers LYVE-1 and VE-cadherin, and upregulated mesenchymal markers including vimentin, fibronectin, N-cadherin and α-smooth muscle actin. Three-dimensional spheroid invasion assays in Matrigel and Boyden chamber migration assays demonstrated that these behaviors depended on SMAD5: silencing SMAD5 with short hairpin RNAs abolished both the invasive growth and the migration induced by vIRF3.

Critical controls established that this metabolic hijacking is separable from the viral life cycle itself. Using bacterial artificial chromosome-based recombination, the team constructed recombinant KSHV clones carrying either wild-type vIRF3, the Δ250-280 deletion mutant, or a revertant restoring the wild-type sequence. The mutant virus replicated normally, produced comparable infectious virions and expressed latent and lytic genes at levels indistinguishable from wild type. Yet when lymphatic endothelial cells were infected with these viruses, only wild-type and revertant infections raised pyruvate kinase activity and acetyl-CoA levels, drove spheroid invasion, and enabled colony formation in soft agar. The phenotypes were reproduced in primary human lymphatic endothelial cells, confirming that the metabolic reprogramming is a genuine consequence of KSHV infection rather than an artifact of immortalized cell lines.

The in vivo consequences were equally striking. In xenograft experiments with immunodeficient NCr-nu/nu mice, cells infected with wild-type or revertant KSHV formed tumors efficiently, whereas cells infected with the vIRF3-MT virus did not. Moreover, mice bearing wild-type-derived tumors developed secondary lesions in the stomach, spleen and liver that displayed hallmark histopathological features of Kaposi’s sarcoma, including extravasated erythrocytes, hemosiderin deposits and spindle-shaped cells positive for the viral latency antigen LANA, vIRF3, the endothelial marker CD31 and the lymphatic lineage factor PROX1. Human vimentin-positive cells at these sites traced back to the implanted infected endothelial cells. No secondary tumors appeared in mice injected with the PKM2-binding-defective virus, directly implicating the vIRF3–PKM2 interaction in disseminated visceral disease.

The study also delivers a potential therapeutic lead. Because amino acids 250 to 260 of vIRF3 were sufficient for PKM2 binding, the researchers designed a short vIRF3-derived peptide, termed VDP, fused to the HIV TAT transduction domain for intracellular delivery. VDP competed with vIRF3 for PKM2 binding, blocked vIRF3-induced PKM2 tetramerization and enzymatic activity, and lowered acetyl-CoA and ATP levels, while a binding-defective mutant peptide, mVDP, did none of these things. Structural modeling with AlphaFold-Multimer and molecular dynamics simulations showed VDP occupying the intersubunit interface of PKM2, and biolayer interferometry measured a dissociation constant of 1.28 micromolar, roughly twenty-nine-fold tighter than the mutant peptide. In chick embryo chorioallantoic membrane assays, TAT-VDP suppressed vIRF3-driven invasion into the endoderm layer, and in xenograft mice the peptide markedly inhibited tumor formation and regressed established tumors without substantial toxicity.

The authors caution that the vIRF3–PKM2–acetyl-CoA–SMAD5 axis still requires validation in patient-derived Kaposi’s sarcoma tissues, which have been difficult to obtain in adequately preserved form, and that their xenograft model does not distinguish true metastatic dissemination from other routes of systemic distribution. Even so, the work provides the most complete mechanistic account to date of how an oncogenic herpesvirus converts a host metabolic enzyme into an engine of tumor progression, linking acetyl-CoA flux through the TCA cycle to epigenetic modification of a developmental transcription factor. By demonstrating that a rationally designed viral peptide can reverse this process in living animals, the study positions PKM2 not merely as a metabolic marker but as an actionable target for treating disseminated Kaposi’s sarcoma.

Subject of Research: KSHV vIRF3–PKM2 interaction, acetyl-CoA metabolic reprogramming and SMAD5 acetylation in Kaposi's sarcoma progression

Article Title: KSHV vIRF3–PKM2 interaction induces SMAD5 acetylation, promoting Kaposi’s sarcoma progression

Article References: Kim, Y. J., Lim, J., Lee, J. J., Kumar, A., Kannappan, S., Zhang, J., Kim, S., Kim, H.-J., Kim, K. K., Toth, Z., Eoh, H., Choi, K.-C., & Lee, H.-R. (2026). KSHV vIRF3–PKM2 interaction induces SMAD5 acetylation, promoting Kaposi’s sarcoma progression. Experimental & Molecular Medicine. https://doi.org/10.1038/s12276-026-01849-2

Image Credits: AI Generated

DOI: 10.1038/s12276-026-01849-2

Keywords: KSHV, Kaposi's sarcoma, vIRF3, PKM2, acetyl-CoA, SMAD5 acetylation, lymphatic endothelial cells, endothelial-to-mesenchymal transition, metabolic reprogramming, p300, TCA cycle, therapeutic peptide